摘要目的 探讨基质金属蛋白酶(MMPs)抗体芯片用于胃癌(GC)诊断的临床应用价值.方法 连续收集2016年11月至2017年3月于我院确诊的12例胃癌(GC)患者和10例胃良性病变(NGD)患者的血清标本组成训练集.应用MMP抗体芯片对上述患者血清中10种分泌型MMPs的浓度同时进行检测.比较其在GC和NGD组间浓度的差异及其用于胃癌诊断的效能.筛选出两组间浓度差异最大的MMP亚型进入下一步研究.继续收集2017年4月至2017年7月于我院确诊的40例GC和38例NGD患者的血清标本组成验证集.应用悬液微珠芯片对上述患者血清中组织型金属蛋白酶抑制剂?4(TIMP?4)的浓度进行检测.比较其在GC和NGD组间浓度的差异及其用于胃癌诊断的效能.结果 训练集中(1)呈高表达(>10 000 pg/ml)的MMP亚型包括:MMP?1、MMP?3、MMP?9和TIMP?1;中表达(1 000~10 000 pg/ml):MMP?8、TIMP?2和TIMP?4;低表达(<1 000 pg/ml):MMP?2、MMP?10和MMP?13.(2)与NGD组相比,除MMP?2、MMP?3和MMP?13外,其余MMP亚型在GC组中的浓度均呈升高趋势.但是仅有TIMP?4差异显著(P<0.05).(3)MMP家族各亚型用于GC诊断的曲线下面积在0.500~0.858之间,联合诊断的AUC为0.925(95% CI :0.782~1.000).其中以TIMP?4最高,为0.858(95% CI :0.692~1.000).当选取2 438.3 pg/ml为最佳临界值时,TIMP?4用于GC诊断的敏感度、特异性和约登指数分别为83.3%、80.0%和0.633.验证集中TIMP?4在GC组和NGD组中的浓度分别为(1 148.0±691.5) pg/ml和(834.3±397.2)pg/ml,组间比较差异有统计学意义(P<0.05);TIMP?4用于GC诊断的曲线下面积为0.628(95% CI :0.498~0.758).结论 血清MMP表达谱的检测有望成为胃癌辅助诊断和筛查的新方法,其中TIMP?4对胃癌的诊断有一定的参考价值,其诊断能力有待于进一步验证.

abstractsObjective To explore the clinical value of matrix metalloproteinase(MMPs)antibody chip in the diagnosis of gastric cancer(GC). Methods Serum samples from 12 patients with gastric cancer (GC)and 10 patients with non?neoplastic gastric disease(NGD)diagnosed in our hospital from November 2016 to March 2017 were collected to form a training set. Human MMP arrays were used to quantitative measurement of 7 major secretory MMPs and 3 tissue inhibitors of metalloproteinases (TIMPs) simultaneously in the above patients. Their concentration differences between GC and NGD groups and their diagnostic efficacy for GC were compared. The MMP subtypes with the greatest concentration difference between the two groups were screened for further study. Serum samples from 40 patients with GC and 38 patients with NGD diagnosed in our hospital from April 2017 to July 2017 were collected to form a validation set. The concentration of TIMP?4 in serum of these patients was detected by suspension microbead chip. Then the levels of TIMP?4 was compared between the GC and NGD groups and its diagnostic value was assessed by ROC curve. Results The MMP subtypes with high expression(>10 000 pg/ml)in the training set included:MMP?1,MMP?3,MMP?9 and TIMP?1;medium expression(1 000?10 000 pg/ml):MMP?8,TIMP?2 and TIMP?4;low expression(<1 000 pg/ml):MMP?2,MMP?10 and MMP?13. Compared with the NGD group, except MMP?2,MMP?3 and MMP?13,other MMPs showed an upward trend in the GC group,but no significant difference. Only TIMP?4 increased significantly in the GC group than NGD group(P<0.05). The area under the curve(AUC)of the MMP family members for the diagnosis of GC ranged from 0.500 to 0.858,and their combined AUC was 0.925(95%CI :0.782?1.000). Particularly,TIMP?4,whose AUC was the highest,was 0.858 (95%CI :0.692?1.000). At the optimal cut?off value of 2 438.3 pg/ml,its sensitivity,specificity and Yoden index for the diagnosis of GC were 83.3%,80% and 0.633,respectively. In the validation set,the concentrations of TIMP?4 in the GC and NGD groups were(1 148.0±691.5)pg/ml and (834.3 ± 397.2) pg/ml, respectively (P<0.05). Its AUC was 0.628 (95%CI : 0.498?0.758). Conclusions The detection of serum MMP expression profile is expected to be a new method for assistant diagnosis and screening of GC. TIMP?4 has certain reference value for the diagnosis of GC,but its diagnostic ability needs further verification.